Special transformer user WeChat group information analysis method based on cloud side-end cooperation

The invention discloses a special transformer user WeChat group information analysis method based on cloud side-end cooperation, and relates to the field of information analysis. The method comprises the following steps: establishing a special transformer user data fusion system based on heterogeneous multilayer cloud edge-end collaboration; performing preprocessing operation and distributed storage on the cloud side information to realize fusion of multilayer data of the cloud side; and performing multi-dimensional analysis on the special transformer user information by using the multi-deep learning fusion model to realize intention recognition of the special transformer user information. According to the technical scheme, through use of multiple layers of edge servers, part of processing tasks in traditional cloud services are subjected to sink processing, the data transmission delay time is shortened, and the computing pressure of a cloud service center is relieved; different neural network processing mode